What is boolean? Is “Boolean Black belt” a good thing? What is Beyond Boolean?
Posted by: Donato in Data Mining, Internet Research, domain names, tags: boolean black belt, broadlean, Broadlook, searchologistThere has been a recent rise of the term “Boolean blackbelt”, while I am not familiar, specifically, with all the people stating to be a a boolean black belt, I wanted to add some perspective.
“Boolean Blackbelt”, may be a great marketing phrase, but it is the equivalent of saying “basic math blackbelt” or “kindergarden green beret”
Why? because Boolean is extremely simple. It is the basis for logic which includes logical operators AND, OR, NOT, and XOR (exclusive OR).
I do know some people out there that would qualify as “search string black belts” and what they do is a combination of understanding boolean logic along with the myriad of Internet sources. They have vast knowledge of the process of search. The magic they bring is not the boolean, it is the business process and understanding. What they do is tweak out the special commands allowed by the search engines, like Google, MS Live. Shally Steckerl and Glenn Gutmacher from Job Machine are examples of masters they are “beyond boolean”.
So please, hold the angry emails, I am not knocking the “black belts” out there, I think they are actually selling their skills short. My goal is to add clarity that being good in boolean may take few minutes. Being good at the complex search commands available in the search engines may take a few days to learn. To know HOW to apply the search fundamentals means you must understand and live with all the sources of information that are available. Search engines, blogs, social networks, etc. To know all that requires total immersion.
What is beyond boolean? Remember, boolean equates to: AND, OR, NOT, OR. Some popular search engine commands include: NEAR, Site, inURL.
(By the way, I just checked the domain: beyondBoolean.com is available!). Who is going to be beyondboolean?
What is beyond boolean? What is missing? Here are some that we have developed at Broadlook. We call it, what else, “broadlean”. The next major versions of Broadlook product engines will be supporting it (some already do). I would like to hear from the searchologists (i prefer that term), what else would you like to see in the broadlean specification?
WARNING!!!!! – The rest of this blog is for the tech geeks out there. If you don’t know your way around a search engine, STOP HERE!
Here are some BROADLEAN commands:
SS: Same sentance as
SP: Same paragraph as
SC: Same concept as
PERSON:
CORP:
DATE:
EVENT:
SS, SP, SC are operators. PERSON, CORP, DATE, EVENT are what we call entity operators.
example usage:
“Project manager” SS PERSON - find all pages that mention the term “project manager” in the same sentance as a person. The value here is that you don’t have to know the person’s name to succeed with your search.
Google, MS Live, Yahoo are all toys when it comes to weeding through results. I am excited to see the technology that comes out over the next few years in terms of targeting results. What fun!
![View Donato Diorio [LION] toplinked.com's profile on LinkedIn](http://www.linkedin.com/img/webpromo/btn_myprofile_160x33.gif){kind=small}
Entries (RSS)
I’m sure you will get some angry emails, anyway. Anyone who is a true Boolean Black Belt and knows enough to teach it well (e.g., the person who actually writes the blog of that name, who I have come to know and respect) has knowledge in applying it far beyond that, i.e., the “total immersion” approach you speak of.
But we need to share and foster quality Boolean training in whatever blogs, classes, discussions, etc., we can. Otherwise, people won’t be ready to learn what comes next, nor be able to take full advantage of the kinds of functionality that Broadlook and others are (and will be) offering — a problem you already suffer with some of your advanced products.
While you are offering the promise of one valuable branch in the evolution of search, we cannot forget to mention another: semantic search. If we can accurately anticipate the user’s wants from context, then we can eliminate much of the human error and learning curve that advanced search currently requires. This field is barely out of its infancy, but VCs apparently are willing to fund wunderkind toddlers: Some interesting companies out there are DeepDyve.com (formerly Infovell) and here in the Boston metro, several companies profiled in this article: http://www.masshightech.com/stories/2009/02/02/weekly6-Smart-analytics-drives-the-semantic-search.html . But the business model so far has been subscription-based for enterprise-only releases. I believe these products will only become great when they’re tested on a wide scale: i.e., when a public search engine adopts it.
So finally, to your question on what Broadlean needs added, I think you need more semantic components. The SC: (same concept as) command appears to come closest. Can you offer that for other categories? For example, if I want events similar to a particular event, because that will likely find me other relevant people, show those, too! For clarity, maybe you should relabel SC to *Similar* Concept as, and keep the “Same as” for the concrete stuff like sentence, paragraph, etc. So now you have the potential for:
- SEVENT (similar industry events, mentioned above)
- SPERSON (similarly-skilled person, same team, etc.) which is what Shally’s Peer Regression Analysis concept focuses on.
- SCORP (similar companies in terms of industry niche for flagship products – which ZoomInfo.com’s free Company Search does much more easily/quickly than Broadlook Profiler and, considering the tradeoffs, fairly well)
- SORG (would be nice if you could apply this to the similar trade association, user group, etc., concept, though maybe you’d fold it under SCORP)
The ultimate result would be an Excel output with each type of data (companies, events, associations/portals, people, etc.) each on their own worksheet tab, and any new results found on an ongoing basis would be added automatically, but also allowing users to manually input new discoveries. With the important exception of some tracking data, this is basically the format of the sourcing research document that JobMachine espouses.